In Free Will and Luck, Mele presents the Zygote Argument, which he takes to be indicative of “the most serious conceptual obstacle for compatibilism (presupposing necessitarianism about the laws of nature)” (2006, 199). The zygote argument centers on the following case of “original design” (hereafter called “Ernie”):

Diana creates a zygote Z in Mary. She combines Z’s atoms as she does because she wants a certain event E to occur 30 years later. From her knowledge of the state of the universe just prior to her creating Z and the laws of nature of her deterministic universe, she deduces that a zygote with precisely Z’s constitution located in Mary will develop into an ideally self-controlled agent who, in 30 years, will judge, on the basis of rational deliberation, that it is best to A and will A on the basis of that judgment, thereby bringing about E. If this agent, Ernie, has any unsheddable values at the time, they play no role in motivating his A-ing. Thirty years later, Ernie is a mentally healthy, ideally self-controlled person who regularly exercises his powers of self-control and has no relevantly compelled or coercively produced attitudes. Furthermore, his beliefs are conducive to informed deliberation about all matters that concern him, and he is a reliable deliberator. So he satisfies a version of my proposed compatibilist sufficient conditions for having freely A-ed (2006, 188).

Mele outlines the following argument against compatibilism with regard to Ernie, although he remains agnostic about the first premise (and hence about the conclusion):

  1. 1.

    Because of the way his zygote was produced in his deterministic universe, Ernie is not a free agent and is not morally responsible for anything.

  2. 2.

    Concerning free action and moral responsibility of the beings into whom the zygotes develop, there is no significant difference between the way Ernie’s zygote comes to exist and the way any normal human zygote comes to exist in a deterministic universe.

  3. 3.

    So determinism precludes free action and moral responsibility (2006, 189).

Premise one is “an assertion about a case” (Mele 2006, 192) and is meant to have intuitive appeal, with the reader perhaps even being agnostic about what led her to find it plausible. Premise two is, Mele argues, consistent with various compatibilist proposals (2006, 192). Taken together, however, the two claims, if true, speak against compatibilist accounts of free will and moral responsibility.

This paper will criticize Mele’s zygote argument. Section one reviews previous approaches to the zygote argument and discusses The Plausibility of Premise One. Section two argues against premise two; Diana’s effective intention to create Ernie’s zygote such that Ernie A-s in 30 years is a significant difference between Ernie and normal agents in a deterministic universe with regard to how their zygotes were created. Finally, I will argue that even if Ernie is somewhat morally responsible for his action, he is still significantly less so than normal twin agents in deterministic worlds because of the presence of Diana’s effective intention.

The Plausibility of Premise One

Several papers have sought to undermine the zygote argument by denying premise one (e.g., Fischer 2011; Kearns 2012). Fischer (2011) takes the critical strategy that Mele (2006) envisions compatibilists would take: he denies premise one and accepts premise two. In other words, Fischer contends (1) that there is no significant difference between Z’s creation and the normal creation of zygotes in deterministic worlds with regard to those resultant beings’ free will and moral responsibility, but (2) that Ernie is indeed acting freely and is morally responsible for his A-ing (and hence for bringing about E). Kearns (2012) approaches the zygote argument via a different route: Kearns is agnostic about premise two–he neither endorses nor rejects the idea that there are no pertinent differences in how Ernie’s zygote and normal zygotes are created in a deterministic universe that bears upon whether the resultant agents freely A-ed and are morally responsible for doing so. Rather, Kearns distinguishes possible readings of premise one and argues that the argument fails regardless of which reading of premise one we assume.

Unlike previous opponents of the zygote argument, I will argue that premise two is false. First, however, it will be helpful to clarify the conception of moral responsibility at issue in the argument via a brief examination of premise one. I initially interpret premise one as the simplified premise that Kearns constructs, that “Ernie is not a free agent and is not morally responsible for anything (in the universe Mele describes)” (2012, 386).

What does it mean to say that an agent is not morally responsible for any given action she performs? In Mele’s discussion of little agents, such as young children, who are plausibly to some small degree morally responsible for some of their behavior, he argues that these agents deserve some credit or blame for their actions (2006, 129–132). Likewise, when proposing how various free will theorists might react to premise one of the zygote argument, Mele links judgments that Ernie is indeed morally responsible for A-ing to judgments that Ernie is blameworthy in cases in which A is a morally bad action (2006, 193). Furthermore, in a paper discussing history-sensitive compatibilism, Mele states the following about his use of the term “morally responsible”:

One occasionally sees in the literature a use of this expression [morally responsible] according to which an agent can correctly be said to be morally responsible for performing some good intentional action for which he deserves no moral credit or some bad intentional action for which he deserves no moral blame. To forestall potential confusion, I report that, as I use “morally responsible,” an agent’s being morally responsible for performing a good intentional action entails that he deserves some moral credit for it and an agent’s being morally responsible for performing a bad intentional action entails that he deserves some moral blame for it (2009b, 465).

Thus, I will assume from here on that premise one of the zygote argument entails that Ernie does not deserve blame (or praise, etc.) for anything he’s done (in the universe that Mele describes).Footnote 1

Keeping in mind that the plausibility of premise one need only rest on the intuitive force of the premise, premise one appears plausible. Ernie does not seem to deserve blame (if A is a morally bad action) or praise (if A is a morally good action) for A-ing or indeed for anything he does.Footnote 2

Denying Premise Two

How might a compatibilist who shares this intuition about Ernie’s not deserving blame or praise–an endorsement of premise one—confront Mele’s zygote argument? There are substantive grounds for rejecting premise two of the zygote argument, thereby rescuing compatibilism from the grip of Ernie.

Premise two of the zygote argument states that: “Concerning free action and moral responsibility of the beings into whom the zygotes develop, there is no significant difference between the way Ernie’s zygote comes to exist and the way any normal human zygote comes to exist in a deterministic universe” (Mele 2006, 189). As noted, unlike Kearns (2012) and Fischer (2011), I deny premise two. Hence, a proper defense of my compatibilist position requires identification and an explanation of the normative significance of some feature present in the causal history of Ernie’s action (but lacking in the causal history of his otherwise twin). To identify such a feature, I will briefly examine which features may be driving intuitions that Ernie is not acting freely and is not morally responsible for A-ing (or any actions he performs) in Mele’s Ernie case that are also notably not present in the case of his otherwise twin.

Intuitions About Ernie

Comparing the creation of Ernie’s zygote to that of normal zygotes in deterministic universes brings to mind one especially salient difference, Diana’s effective intention to create Ernie’s zygote such that Ernie A-s in 30 years (and the lack of such effective intentions of creators in normal situations). “Effective intention to A” in a deterministic universe at least entails that A occurs. One can also adopt a stronger reading of “effective intention to A” in a deterministic universe: if an agent S has an effective intention to A in a deterministic universe, this entails that the S intentionally performs that intended action A. So, to illustrate, to say that Diana has an effective intention to create Ernie’s zygote such that Ernie A-s in 30 years entails, on the stronger reading, that Diana, acting on that intention, intentionally creates Ernie’s zygote such that he A-s in 30 years (and he does A in 30 years). For short, Diana’s intention will be abbreviated as “the effective intention that Ernie A-s.”

Fischer (2011) argues that the intentions of the creators of the zygote are irrelevant to intuitions regarding Ernie’s desert of blame or praise, but this relies on the assumption that all intentions are created equal. Like Todd (2012), I contend that a feature of Ernie’s case that is not present in Fischer’s cases of the otherwise twins of Ernie—the presence and effectiveness of Diana’s specific intention to create Ernie’s zygote such that he A-s—is what actually accounts for the intuitive force of premise 1. Thus, in the case of at least some who have the intuition that Ernie isn’t morally responsible, Diana’s role in the story helps to explain why they have that intuition. But is there a way to support the claim that this intuition is tracking a relevant factor for determining moral responsibility?

Principle Behind Denying Premise Two

Here I will argue that the effective intention of Ernie’s creator marks a significant difference between how Ernie’s zygote comes to exist and normal zygotes come to exist in deterministic worlds, one that affects whether the resultant agent is morally responsible for his action. As a first pass, we can support this claim with the following thesis (Further support for the claim that Ernie deserves no blame or praise for A-ing will be addressed later):

T: S is less deserving of blame or praise for A-ing than she would be otherwise if (1) another agent G effectively intends that S A-s and (2) S did not intentionally bring it about that G intends that S A-s.

Note that T refers to another agent’s effective intention, as opposed to the narrower class of a creator’s effective intention. This move is needed to cover variants of the Ernie case, such as a test-tube Ernie case in which Ernie’s parents intend for Ernie to A, thereby bringing about event E, and so consult a genius fertility specialist (who has the same knowledge of the state of the universe and laws of nature as Diana does). The second condition, that S did not intentionally bring it about that G intends that S A-s, is needed so that an agent does not get off the hook for A-ing when he is responsible for the other agent’s effective intention that he A-s.Footnote 3

Thesis T captures the intuition that in cases of extreme brainwashing, drugging, and hypnosis, the agent is not acting freely and is not morally responsible for her action. Consider the following pair of agents and accompanying claims about moral responsibility from Mele 2009a:

Van, a normal man, got drunk at a party and then tried to drive home. He was so drunk that he did not realize he was impaired. No one tricked Van into drinking alcohol, no one forced him to drink, he is knowledgeable about the effects of alcohol, and so on. Owing to his drunkenness, he drove into and killed a pedestrian he did not see. A plausible judgment about Van is that, other things being equal, he is morally responsible for killing the pedestrian.

Ike was just as drunk as Van when Ike tried to drive home, so drunk that he did not realize he was impaired. Owing to his drunkenness, Ike drove into and killed a pedestrian he did not see. However, Ike was force-fed alcohol and placed in the driver’s seat of his car. It is very plausible that he is not morally responsible for killing the pedestrian (162).

If these judgments about Van and Ike’s moral responsibility for killing the pedestrian are correct, assuming that both share the same psychological constitution and control over their actions at the time of the killing, some differential feature of Ike’s history in contrast to Van’s explains why he lacks moral responsibility for the killing (Mele 2009a). Here a good candidate for such a feature is another intentional agent–the culprit Calvin who force-fed Ike and placed him behind the wheel. Hence, if Calvin intended to bring it about that Ike gets in an accident, or even more specifically kills a pedestrian, this aspect of Ike’s history is relevant for determining Ike’s moral responsibility. In such cases, the effective intention of the other agent G brings about the agent S’s A-ing–either via S’s deliberation or via bypassing S’s deliberation. Hence, G’s effective intention is a key part of the causal history of S’s action that makes it the case that S does not act freely and does not deserve any blame for A-ing.Footnote 4

Thesis T also coheres with intuitions about moral responsibility in related cases in which the agent acts at least partially due to the effective intention of another, but the influence of that intention is not so strong as to render it inappropriate to blame or praise the agent for that action. In such cases the agent may be at least somewhat deserving of blame and praise for her action. For instance, imagine that a foe of mine knows that I have given up eating chocolate for Lent. This foe, without my intentionally bringing it about, intends to bring it about that I eat chocolate despite my pledge. During Lent, she eats my favorite chocolate, a Cadbury bar, in front of me. She succeeds; I give in to temptation and eat a Cadbury bar during Lent.Footnote 5 Or alternatively, imagine a case in which a cult leader, who desires that a certain person, Victor, be killed, intends to bring it about that a member, Kirk, kills Victor and then executes a plan of influence on Kirk so that that Kirk kills Victor. In assessing my eating the Cadbury bar and Kirk’s killing Victor, a reader may feel that both agents deserve at least some blame for their actions. According to T, both Kirk and I are deserving of some blame for our actions but deserve less blame than an agent that is not subject to such an effective intention of another in a deterministic universe.

Another (likely) common intuition about moral responsibility is that if the influence of another’s effective intention that an agent A-s is weak enough, then the agent’s moral responsibility for A-ing is not significantly diminished. T is consistent with this fact: T does not entail that all agents who act because of another’s effective intention that they so act have significantly mitigated moral responsibility. To see this, consider two variants of the Kirk case: (K2) Suppose that Kirk is a violent man who already hates Victor and is seriously considering killing Victor but has not yet decided to do so. Kirk’s friend also hates Victor and would like to see Victor killed. Kirk’s friend knows that Kirk hates Victor and is prone to violent acts, so his friend merely suggests to Kirk that he kills Victor, intending that Kirk do so. Because of his friend’s suggestion, Kirk kills Victor. Compare this case to a similar case (K3) in which Kirk decides to and does kill Victor before his friend suggests that he does so. In both cases, provided Kirk satisfies certain conditions, Kirk deserves a lot of blame for killing Victor. However, note that this is still consistent with there being a negligible–hardly noticeable—difference in the degree to which Kirk deserves blame in the first case versus Kirk in the second. What tipped the scales towards Kirk’s killing Victor in the first case was his friend’s intervention; in the second case, he decided to kill Victor on his own. T claims that this differential causal history of their actions matters, although it may result in practically the same degree of deserved blame when the effective intention is a weak one. “Defense of Principle T*” will provide a framework for explaining the difference in degree of responsibility in such pairs of cases.Footnote 6

An Objection to Thesis T

Before discussing the explicit role of T in the rejection of premise two in the zygote argument, a potential counterexample to T should be addressed. In the type of case at issue, the other agent G has an effective intention that the agent S A-s but G needn’t intervene to bring it about that S A-s. Perhaps G only needs to refrain from intervening to make her intention effective. For example, suppose that Tom notices that the safe in the main office at work has been left open. Tom knows that a co-worker Sam has been eyeing the safe with an intention to steal the cash inside and further that Sam will be in the main office soon. However, Tom wants Sam to steal the cash in the safe–perhaps to get back at his boss. So Tom doesn’t lock the safe. Sure enough Sam decides to and does steal the cash in the safe.Footnote 7 A variant of the Kirk case has a similar structure: Instead of Kirk’s friend suggesting that Kirk kill Victor as in K2, his friend knows that Kirk will decide not to kill Victor and will not in fact kill him only if he talks Kirk out of it. As in the previous case, Kirk’s friend intends that Kirk kills Victor, so he refrains from talking Kirk out of doing so. Kirk kills Victor.

Intuitively, it looks as though Sam and Kirk are just as morally responsible as they would have been had they decided to act and so acted without any other agent’s intending them to do so. After all, if the other agent–Tom and Kirk’s friend–hadn’t been a feature of the case, Sam and Kirk would have performed the same action. To put this type of scenario in the context of the zygote argument, imagine that Diana deduces on the basis of her knowledge of the laws and nature and state of the universe in her deterministic world that the zygote Z as it has already been constituted entails that Ernie will A thereby bringing about E. Because Diana wants E to occur she refrains from interfering in Z’s development. It is plausible that Ernie deserves just as much blame (or praise) for A-ing as he would have otherwise if Diana did not intend for Ernie to A.Footnote 8

This type of problematic case for T illuminates an important point about the original Ernie case: it is not just the fact that Diana effectively intends that Ernie A-s that marks the difference between the development of Ernie’s zygote and that of his otherwise twin. Rather, more accurately, Diana’s intervention–her so designing Ernie’s zygote so that he A-s—in conjunction with her effective intention is the relevant difference. And, as the aforementioned examples attest, the presence of another agent that effectively intends that S A-s and so intervenes to bring it about that S A-s is a partial determinant of how blameworthy or praiseworthy S is for A-ing. Thus, the generalized thesis that supports a rejection of premise two in the zygote argument should be modified as follows:

T*: S is less deserving of blame or praise for A-ing than she would be otherwise if (1) another agent G effectively intends that S A-s, (2) G brings it about that S A-s via intervention, and (3) S did not intentionally bring it about that G intends that S A-s.

Here intervention includes overt actions and the consequences of overt actions but excludes omissions, such as refraining to act.

T* and Premise Two

T* provides substantive grounds for criticizing the zygote argument. Premise two of the zygote argument is that there is no significant difference between the way that Ernie’s zygote was created and the way that normal human zygotes are created in a deterministic universe that bears upon the resultant agents’ free will and moral responsibility. According to T*, given the lack of another agent’s effective intention and intervention to bring about that Ernie’s (otherwise) twin A-s, there is such a difference between Ernie and his twin. Namely, even if Ernie is somewhat deserving of blame for A-ing (and so satisfies minimal threshold conditions for acting freely), he is less so than his twin is. Of course, T* is not sufficient to reject premise two. Premise two claims that there is no such significant difference, and T* by itself does not secure a significant difference between Ernie and his normal twin. As acknowledged above, in some cases another’s effective intention that an agent A-s and intervention only mitigates degree of blame slightly–in a non-significant way. Thus, more work needs to be done here to show that the creation of Ernie’s zygote and of his otherwise twin’s zygote is a significant difference that affects their desert of blame or praise for their actions.

To establish that Diana’s effective intention that Ernie A-s and her intervention to bring it about mark a significant difference between Ernie and his otherwise twin, one need only look to the details of Mele’s case. For the remainder of the paper, the focus will be on cases in which another agent effectively intends that S A-s and intervenes in some way to bring it about that S A-s. Hence, any discussion of another agent’s effective intention that an agent A-s should be understood as effective via intervention, but will not be explicitly referred to as such for the sake of brevity. The effective intention of Kirk’s friend that Kirk kills Victor in K2 did not significantly affect Kirk’s blameworthiness. Now compare that intention to Diana’s effective intention for Ernie. Although both cases feature another agent’s effective intention that some agent performs an action, there is an important difference: Whereas Kirk’s friend’s intention encompasses one action that Kirk will perform, Diana’s intention casts a much wider scope–the way in which Diana designs Ernie’s zygote so that he A-s in 30 years dictates all of his actions. Thus, Diana’s intentional action of so constituting Ernie at his creation affects Ernie’s agency in a profound way and so has prime bearing on whether he acts freely or is morally responsible. This type of interference is missing from the case of Ernie’s otherwise twin. Given the similarity between Ernie and normal agents in a deterministic universe in all other respects, then, there is a significant difference in the causal history of their agency. Thus, Diana’s effective intention that Ernie A-s in 30 years is a significant difference between the way Ernie’s zygote and the zygotes of normal agents come to exist in a deterministic universe that is relevant for assessing whether such agents are morally responsible for their actions. Premise two of the zygote argument is false.

The preceding explanation of how T* may be employed to reject premise two assumed that Ernie is somewhat morally responsible (although significantly less so than his otherwise twin). This would also deny premise one. However, any compatibilist that accepts premise one of the zygote argument can also rely upon the principle T* in rejecting premise two of the zygote argument. T* is consistent with the claim that Ernie is not a free agent and is not morally responsible for anything (in the universe that Mele describes).Footnote 9 A justified rejection of premise two suffices to undermine the zygote argument, regardless of the truth-value of premise one.

Defense of Principle T*

Thus far, I’ve argued that even if a compatibilist grants premise one of the zygote argument, there are grounds for rejecting premise two–namely that Diana’s effective intention that Ernie A-s and her so constituting Ernie’s zygote marks a significant difference between the way that Ernie’s zygote comes to exist and the way a normal human zygote comes to exist in a deterministic universe in reference to their moral responsibility for A-ing. A plausible principle supports this claim, termed T*: S is less deserving of blame or praise for A-ing than she would be otherwise if (1) another agent G effectively intends that S A-s, (2) G brings it about that S A-s via intervention, and (3) S did not intentionally bring it about that G intends that S A-s. This last section offers an argument for the tenability of T*.

In a recent paper arguing for a reasons-responsive account of moral responsibility that allows for degrees of responsibility, Coates and Swenson (2012) elucidate why it may be justified to say that two agents who perform the same action in similar circumstances can be morally responsible and deserving of blame (or praise) to different degrees. They contend that among the class of agents who satisfy the minimal threshold conditions for moral responsibility in A-ing, the degree to which those agents are able to recognize and act on reasons for acting corresponds to the degree to which they are morally responsible for A-ing. Coates and Swenson build their theory upon Fischer and Ravizza (1998)’s reasons-responsive theory of moral responsibility, taking Fischer and Ravizza’s theory as specifying at least a necessary threshold condition for ascribing moral responsibility to and blaming (or praising) an agent. For Fischer (and Ravizza) an agent is morally responsible–an apt target for the reactive attitudes—only if she satisfies guidance control in A-ing: That is, the process that actually leads to action (the actual-sequence mechanism) is appropriately reasons-responsive. By appropriately reasons-responsive, Fischer and Ravizza mean that that mechanism displays an adequate capacity both to recognize the reasons for acting and to act on those reasons in the given scenario. Here reasons-responsiveness codifies and precisifies the common thought that agents who qualify as a target of the reactive attitudes for A-ing must display adequate capacities both to recognize the reasons to act, including moral reasons, and to act on those reasons in the given situation.Footnote 10

First, to illustrate how two agents who perform the same action in similar circumstances can be morally responsible and deserving of blame to different degrees, Coates and Swenson (2012) offer the examples of Marcia and Thomas. Both Marcia and Thomas promised to pick a friend up from the airport and failed to show up at the agreed-upon meeting time. Coates and Swenson note that if you were that friend, you would likely hold Marcia or Thomas responsible and blame Marcia or Thomas for failing to pick you up. You would also likely resent her or him. But now, Coates and Swenson add, Marcia tells you that she is suffering from depression. She discloses that because of her depression she found it extremely difficult (although possible) to leave her apartment and so didn’t leave, even though she recognized that she ought to leave to pick you up as promised. Coates and Swenson also explain why Thomas failed to meet you at the airport: Thomas, they say, tells you that his favorite movie was on TV at the time he was to pick you up, and he wanted to watch it. So, Thomas did not leave his apartment to pick you up from the airport so that he could watch the movie, despite recognizing that he ought to pick you up as promised. Coates and Swenson stipulate that both Marcia and Thomas satisfy Fischer and Ravizza’s minimal threshold conditions for moral responsibility. They note that while Marcia’s depression would mitigate the degree to which you blame and resent her for failing to meet you, Thomas’s explanation does not–re-watching a favorite movie does not outweigh an obligation to uphold a promise to shuttle a friend from the airport. So, Coates and Swenson point out that in common cases of agency we do hold agents morally responsible and blame (or praise) agents to different degrees for similar actions on the basis of certain factors in the causal history leading to their action.

Coates and Swenson then explain degrees of moral responsibility and blame in virtue of how reasons-responsive an agent’s actual-sequence mechanism that issues in A-ing is–the degree to which that mechanism displays the modal properties of the capacity to recognize reasons and the capacity to react to reasons. Specifically, they contend that

how responsible and blameworthy an agent is for her action can depend on the modal properties of the actual-sequence moderately reasons-responsive mechanism that issues in her action. If those modal properties are made true by comparatively similar worlds, we argue that the agent in question enjoys a high degree of responsibility. Conversely, if those modal properties are made true by comparatively dissimilar worlds, the agent in question is less responsible (2012).

Here Coates and Swenson understand the relation of “comparative similarity” between worlds in terms of nearness of worlds, including but not limited to considerations of shared true propositions across worlds. The theory assesses the degree of moral responsibility of agents who already satisfy the minimal threshold conditions for moral responsibility in terms of the nearest possible world to the actual world in which it is true that if there was a sufficient reason to act other than the agent actually did, the agent acting via her actual sequence mechanism would recognize that reason and act otherwise for that reason.Footnote 11

Consider this contention with regard to Marcia and Thomas: We intuitively think, assert Coates and Swenson, that Marcia’s depression mitigates her moral responsibility for failing to pick up her friend. Why? Because of her depression, it is harder for Marcia than for Thomas, who is not depressed, to react to a sufficient reason to pick up a friend at the airport. Indeed, the proposed theory explains this fact: In terms of possible worlds, the nearest possible world Wp in which there is there is a sufficient reason for Marcia to pick up her friend at the airport and she does pick up her friend for that reason is comparatively dissimilar to the actual world–presumably in such a world Wp Marcia is not depressed or there is a strong reason (compared to the actual reason) to meet her friend. These worlds feature major differences from the actual world in terms of the set of true propositions in the two worlds. In comparison, the nearest possible world Wp in which there is a sufficient reason for Thomas to pick up his friend at the airport and he does for that reason is near–comparatively similar—to the actual world. A world with only minor differences from the actual world, say one in which the movie was playing on TV a little later or earlier, is a world in which Thomas meets his friend at the appointed time at the airport. So, the relevant Wp is less accessible to Marcia than it is to Thomas (Coates and Swenson 2012).

Returning to the discussion of Mele’s zygote argument, we can now justify the adoption of principle T* using the machinery of Coates and Swenson (2012)’s theory of degrees of moral responsibility. Recall that T* is the thesis that S is less deserving of blame or praise for A-ing than she would be otherwise if (1) another agent G effectively intends that S A-s, (2) G brings it about that S A-s via intervention, and (3) S did not intentionally bring it about that G intends that S A-s. To test T* in light of Coates and Swenson’s theory, consider the following cases: Two agents, Hanna and Anna, both steal an identical bracelet from Betty’s Boutique because each thinks the bracelet is beautiful. Assume that both Hanna and Anna satisfy the minimal conditions for thin moral responsibility. But, now, we come to find out that Hanna has been raised by her mother Helga to value the activity of shoplifting: Helga intends for her daughter to develop into an agent who shoplifts items as often as possible. When Hanna was a child, Helga would take Hanna to shops and make her practice shoplifting. Because of this training, Hanna has developed into an agent who feels the urge to shoplift on a regular basis. Hanna does not, however, have a literally irresistible urge to shoplift; she can easily refrain from doing so if there is a cop present in the store or if she is with an acquaintance who thinks shoplifting is morally wrong. In contrast, Anna has no such history of childhood training. In fact, she has never shoplifted before. She has felt an urge to do so once or twice before, and she decided to give in and shoplift only this once.

Intuitively, Hanna deserves less blame for shoplifting than Anna does. It is more difficult for Hanna to refrain from stealing the bracelet than it is for Anna because of the causal history that leads to Hanna stealing the bracelet: The path that led Hanna to steal the bracelet traces back to Helga’s wide-scope effective intention that Hanna become an agent who shoplifts often and Helga’s program of intentionally raising Hanna so that she becomes a recidivist shoplifter. The causal history of Anna’s stealing the bracelet, in comparison, lacks any other agent’s such effective intention (that she shoplifts) and the corresponding intervening to bring it about that Anna shoplifts. In terms of the comparative similarity of worlds in which the agent has a sufficient reason to refrain from stealing and does refrain for that reason, the nearest world in which Hanna refrains from shoplifting is comparatively dissimilar–such a world either lacks her mother’s intentionally molding her into an agent who steals often or contains a strong reason to refrain (e.g., a cop standing next to her). The nearest world in which Anna refrains from stealing the bracelet, however, is comparatively similar (e.g., one in which she dismisses such urges as silly or one in which she thinks the bracelet is not quite as beautiful). Thus Hanna deserves significantly less blame than Anna does for stealing the bracelet. Indeed, this result matches the verdict we get when we apply T* to the cases of Hanna and Anna. When an agent is acting under the influence of another agent’s effective intention that the agent so act (again, effective via intervention), the nearest possible world in which the agent acts otherwise will be comparatively dissimilar to the actual world; whereas, the nearest possible world in which a twin of that agent in all other respects (that is, minus the third party’s effective intention and intervention) acts otherwise will be comparatively similar.Footnote 12 We can apply this explanation of T* directly to Mele’s Ernie case: the nearest possible world in which Ernie does otherwise than A (or any action he actually performs) is comparatively dissimilar to the actual world; the nearest possible world in which an otherwise twin of Ernie in a deterministic world does otherwise than A is comparatively similar to the actual world. Hence, Ernie deserves significantly less blame or praise than his otherwise twin does for A-ing.Footnote 13

This paper argues that a compatibilist responding to the zygote argument can undermine the argument even if she accepts premise one: Premise two of the zygote argument states that there is no relevant difference between the way Ernie came into being and the way that normal agents in deterministic worlds come into being that bears upon whether the agents deserve moral blame (or praise) for A-ing. However, Diana’s effective intention that Ernie A-s and her so constituting Ernie’s zygote affects Ernie’s desert of blame or praise for A-ing in a way that denies premise two of the zygote argument. There is not similar justification for the actual application of reactive attitudes (and blame and praise) to Ernie and his otherwise twin in deterministic worlds, so the zygote argument does not succeed in establishing the conclusion that the truth of the thesis of determinism rules out free will and moral responsibility. This position is supported with the general principle that an agent is less deserving of blame or praise for her action than she would be otherwise if another agent effectively intends via some intervention that she so acts and she did not intentionally bring it about that the other agent has this effective intention.